LiDAR 기반 측위시스템의 철도환경 적용성 및 주행속도에 따른 성능 분석

초록

This study evaluated the applicability and technical limitations of LiDAR-based localization for railway environments through vehicle driving tests. The tests were conducted on an approximately 1,027 m road section adjacent to the Osong railway test track. A Velodyne VLP-16 Hi-Res LiDAR and an RTK-GPS reference positioning system were installed on the test vehicle. The vehicle positions were estimated by matching real-time LiDAR data with a pre-built three-dimensional point cloud map using NDT and compared with the RTK-GPS reference positions at speeds from 10 km/h to approximately 108 km/h. LiDAR localization was continuously maintained without registration failure up to approximately 108 km/h, with a maximum position error of approximately 2.5 m. The localization performance was relatively stable up to 60 km/h, whereas the errors and variability increased with vehicle speed because of motion distortion and point cloud registration errors. These results highlight the potential of LiDAR map-based localization for low- and medium-speed railway environments and GNSS-denied areas. For high-speed applications, sensor fusion with IMU, RTK-GNSS, and encoder information is required to improve motion compensation and localization stability.

키워드